The necessity of a SECONDARY CIRCUITS in ME EQUIPMENT (IEC 60601-1)

[DENich]

Hello,
As far as I know, there is no direct requirement in IEC 60601-1 that the ME EQUIPMENT has a SECONDARY CIRCUIT. Am I right?

To be more specific let's imagine the flyback converter with D2Dn LED array as a load (see the picture). Capacity of C1 and the charge voltage are quite big, say, 10mF and 1000V. The medical effect is accomplishing by applying the light from D2Dn to a skin or something like that. The device is CLASS I and it has PROTECTIVELY EARTHED ENCLOSURE. Whole the enclosure is an ACCESSIBLE PART. There are clearances and creepages between the inside circuit and the enclosure, which provide 1MOP . There aren't MOPs between different sides of transformer TR1.

Does such ME DEVICE have the right to exist? Would there be any special requirements to the power cord?

 

[Peter Selvey]

In theory it's OK provided that the 1MOP is applied based on the voltage in the flyback converter e.g. 1000V.

The catch is that you would then have to apply that to all the insulation in the mains parts as well, and they are not designed for 1000V working voltage (they are typically designed for 250V working voltage). This includes not only the parts in the device, but all mains parts in the wall supply and any other device that's connected to the same mains circuit.

So it's not going to work. You need double insulation in the transformer to prevent the 1000V folding back on to the mains part.

Now ... 1000V may seem high but if it is a secondary circuit and highly regulated it may be that the requirements are no so difficult to meet.

 

[DENich]

Hello Selvey,
Thanks a lot for the response.

I understand your logic, but I have some doubts. I've seen plenty of power supplies with PFC circuit in their MAINS PART. Usually the voltage on the capacitors of PFC is about 400-450Vdc which is significantly higher than the amplitude of the 230Vac MAINS. IEC 60601-1 (paragraph 4.10.2) argues that that SUPPLY MAINS shall be assumed to belong to overvoltage category II for mains transients. For the 230Vac mains it means that mains transient voltage can be up to 2500Vpeak.

What are you think about it?

 

[Peter Selvey]

There's a few things there. Air clearances are based on overvoltage transients on the mains (things like suddenly turning off an inductive load, fault currents with inductive coupling in cables in the same tray) that can be a few thousand volts. These can also get amplified by transformers in a device and also appear on top of the normal working voltage. So, if you have 1000V output via 1:4 transformer, 2500Vp transient occurs in the primary it could be 11000Vp in the secondary. Or not. It's just a bunch of worst case assumptions in the end, and chances are you would not get 11000Vp but something in between.

Creepages distances and dielectric however are based on working voltages. Here the difference between 1000V and 250V is quite big.

The PFC is a bit complicated, due to the diode rectifier which feeds the main 450Vdc caps, there is a complex relation with the voltage with respect to earth, which means the actual voltage is in the order of 280Vrms, which is not that much higher. The test agencies (should) also verify the insulation at that point in the circuit to earth using both the actual peak (which can often be as high as 800Vp) and rms. Finally, via neutral (which is earthed in the supply) the PFC is earth referenced and both the 0V and 450V of the PFC are very different potential to earth (even the 0V line is about 130Vrms different to earth), if any point in that circuit were to short to earth, the fuse would blow. That means it is not possible for the 800Vp/280Vrms to fold back to the mains supply.

Your case is bit different: you have a functionally isolated 1000V output. Since it is only functionally isolated, we can short the 0V to ground, and then we have 1000V would be stressed in the transformer. But since the transformer is functional only, we can short that too, and you end up with a direct path between mains and earth, which is not good since this would occur in normal condition (shorting functional insulation is considered normal condition). So it seems from that point you need at least 1 MOP at 1000V.

My theory about the folding back on the mains needs more thought - it seems likely that a fuse would blow somehow same as for PFC. If you could prove that the fuse will blow, then maybe 1MOP is enough.

 

[DENich]

First of all, thanks a lot for your detailed reply. I really appreciate it.
I'd like to clarify a few things.

1. "Air clearances are based on overvoltage transients" Do you mean that exact values of required clearances provided in 60601-1 are based not only on an actual working voltage but on possible overvoltage transients too (i.e. if they were based only on an actual working voltage, we would have lesser values in the tables of the standard) ? Because 60601-1 states that choosing a proper clearance is based on Vpeak or Vdc of actual working voltage or on the working voltage which can appear between two parts of a circuit under consideration.

2. Speaking about the third paragraph of your post, I agree that the fuses would blow up in case of failure. But how quick they would blow? I can imagine a scenario when they would blow only after 10ms (half sine wave). Is it safe enough? You are speaking about 800Vpeak (I can think of only 450Vpeak on the main capacitors of a PFC in reference to a neutral, but let it bee). Usually, fuses have rated voltage around 400...500V. Will they safely brake the circuit with greater voltages (I mean without an arc)?
Actually, 450V PFC is just a common practice but we can develop a PFC (or another similar circuit) with higher voltage. Where is the limit of voltage?
Moreover, fuses are not listed in 60601-1 as MOP. Even if we condsider a fuse as MOP we will still need to have another MOP.

3."Since it is only functionally isolated, we can short the 0V to ground"
Why? From what it is derived?

Curiouser and Curiouser.. ((C) Alice In Wonderland). My head is gonna blow up like those fuses))

 

[Peter Selvey]

Q1: Yes, the breakdown in air is about 3kV/mm so even with a large safety factor a gap of 0.5mm would be more than enough to support 240Vrms with negligible risk of breakdown. But if you actually monitor mains voltage with an oscilloscope you will find there are all sorts of tiny transients sitting on top of the mains voltage. Most of these are fairly small (say a few hundred volts) but occasionally they can be large. The working voltage influences this because the designers of the mains system have different design aspects, calculations, limits, protective devices. For example the transients in Australia (230V system) will be different to the 110V system in Japan. International work has harmonized these into different groups based on "overvoltage category" and the mains supply.

Q2: Fuses are part of the MOP structure. In a classic basic insulation + earth (Class I device), the basic insulation is 1MOP and earthing is 1MOP. But it only works if there is a fuse that blows if the basic insulation fails. There is a risk that in the time it takes for the fuse to blow (1 cycle), the earthed part could be raised to dangerous potential. However, the risk of electric shock from a single cycle is actually quite low, the risk of fibrillation comes the current messing up a particular part of the cardiac cycle (repolarization? it's been too long) which is only there for a short time. If you are subjected to a long term shock you are getting 50 or 60 hits a second, there is an increased chance of things lining up. But in truth most people survive electric shocks without fibrillation (I can attest to). Also, there is the additional probability that the person is touching the earth frame with one hand while standing with bare feet on a wet concrete floor, plus the voltage on the frame won't be full mains ... coupled with the 1 cycle factor makes it all an acceptably long shot.
In the case of PFC fail, the fuses won't see 450V, they will see thousands of volts. When a fuse breaks it is opening an inductive circuit which can create voltages far in excess of the supply (see point #1 above). There's also a large amount of energy released. This is why mains fuses should be the "H" type which are filled with sand to absorb the energy. If use the wrong type they can explode.

Q3: It's specified in the standard, (Clause 8.1a from memory ...) any insulation which is not meeting the requirements of Clause 8 can be shorted as "normal condition". Consider a system which has insulation A, B and C in series. But insulation A is just a decorative decal and not designed/intended for safety. So, we short (remove) insulation A and perform the tests on B, C only. However, if shorting a part reliably blows a fuse, it is considered to be fail safe situation.

 

[DENich]

"any insulation which is not meeting the requirements of Clause 8 can be shorted as "normal condition" " (post #6) - this statement is understandable but you wrote "Since it is only functionally isolated, we can short the 0V to ground" (post #4) . I assumed that in my scheme (post #1) there is 1 MOP between any point of the inside circuit and the protectively earthed enclosure. So how we can short the 0V to ground if there is 1 MOP?

 

[Peter Selvey]

My mistake, result of too many late nights. So, if the 1MOP fails and the fuse blows reliably, it's OK. There's still a catch in that there might be a large number of tests: the transformer is 0MOP so it can be shorted as "normal condition" in a few places, and the 1MOP on the output side could also have a few locations. But in principle it is OK.

 

[DENich]

Thank you Peter,
I appreciate your help!
So, to sum up, the scheme in the post #1 has the right to exist in case of the fuse blows reliably If 1MOP fails.